Skip to contentSkip to conversation

scsi: dpt_i2o: fix a potential use-after-free in __adpt_reset()

PATCH In review Hang Zhang 1 participant 4 messages Dec 27, 2022 – Feb 7, 2023 2 messages not in the archive SCSI

lore
Share & tools
lore

Hang Zhang

author
#0 lore raw

PATCH 1/1 scsi: dpt_i2o: fix a potential use-after-free in __adpt_reset()

__adpt_reset() invokes adpt_hba_reset(), which can free "pHba" on error paths and return an negative error code in those situations. The problem is that "pHba" is from the global pointer "cmd->device->host->hostdata[0]", regardless of the possible free of "pHba", that original global pointer is never nullified and thus may become a dangling pointer and be dereferenced later, potentially causing a use-after-free.

Fix the issue by nullifying "cmd->device->host->hostdata[0]" if adpt_hba_reset() returns a negative error code to __adpt_reset(), which indicates the free of "pHba". Also add a NULL check before any dereference of "pHba", or the aliased global pointer. Note that the similar NULL check already exists at other places, like in adpt_queue_lck().

---
 drivers/scsi/dpt_i2o.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/scsi/dpt_i2o.c +4 -0
index 2e9155ba7408..9827517a1898 100644
@@ -753,6 +753,9 @@ static int __adpt_reset(struct scsi_cmnd* cmd)
 	char name[32];
 
 	pHba = (adpt_hba*)cmd->device->host->hostdata[0];
+	if (!pHba) {
+		return FAILED;
+	}
 	strncpy(name, pHba->name, sizeof(name));
 	printk(KERN_WARNING"%s: Hba Reset: scsi id %d: tid: %d\n", name, cmd->device->channel, pHba->channel[cmd->device->channel].tid);
 	rcode =  adpt_hba_reset(pHba);
@@ -760,6 +763,7 @@ static int __adpt_reset(struct scsi_cmnd* cmd)
 		printk(KERN_WARNING"%s: HBA reset complete\n", name);
 		return SUCCESS;
 	} else {
+		cmd->device->host->hostdata[0] = NULL;
 		printk(KERN_WARNING"%s: HBA reset failed (%x)\n", name, rcode);
 		return FAILED;
 	}
Signature · 1 line
2.39.0
author
#1 lore raw
Quoted 44 lines · Hang Zhang

__adpt_reset() invokes adpt_hba_reset(), which can free "pHba" on error paths and return an negative error code in those situations. The problem is that "pHba" is from the global pointer "cmd->device->host->hostdata[0]", regardless of the possible free of "pHba", that original global pointer is never nullified and thus may become a dangling pointer and be dereferenced later, potentially causing a use-after-free.

Fix the issue by nullifying "cmd->device->host->hostdata[0]" if adpt_hba_reset() returns a negative error code to __adpt_reset(), which indicates the free of "pHba". Also add a NULL check before any dereference of "pHba", or the aliased global pointer. Note that the similar NULL check already exists at other places, like in adpt_queue_lck().

---
 drivers/scsi/dpt_i2o.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/scsi/dpt_i2o.c +4 -1
index 2e9155ba7408..9827517a1898 100644
@@ -753,6 +753,9 @@ static int __adpt_reset(struct scsi_cmnd* cmd)
        char name[32];
 
        pHba = (adpt_hba*)cmd->device->host->hostdata[0];
+       if (!pHba) {
+               return FAILED;
+       }
        strncpy(name, pHba->name, sizeof(name));
        printk(KERN_WARNING"%s: Hba Reset: scsi id %d: tid: %d\n", name, cmd->device->channel, pHba->channel[cmd->device->channel].tid);
        rcode =  adpt_hba_reset(pHba);
@@ -760,6 +763,7 @@ static int __adpt_reset(struct scsi_cmnd* cmd)
                printk(KERN_WARNING"%s: HBA reset complete\n", name);
                return SUCCESS;
        } else {
+               cmd->device->host->hostdata[0] = NULL;
                printk(KERN_WARNING"%s: HBA reset failed (%x)\n", name, rcode);
                return FAILED;
        }
--
Quoted 2 lines

2.39.0

Hi James and Martin, could we get some feedback from you regarding this patch? We want to clarify that this issue was originally detected by our static analyzer, so we do not have PoC or dynamic execution traces. However we do have a careful manual code review to confirm the issue, and we think that it seems apparent that the global pointer should be cleared after its points-to object gets freed (and checked before use), which is not the case in the current __adpt_reset() function. Please feel free to let us know if we missed anything here, thank you very much!

Best,
Hang

A message that is not in the archive (sent off-list or not yet received): [email protected]

author
#3 lore raw
Quoted 9 lines · Tomas Henzl

Hi Hang,

with b04e75a4a8a8 scsi: dpt_i2o: Remove obsolete driver was the driver removed from kernel.

Regards,
Tomas

Hi Tomas, thank you very much for pointing this out! I'm wondering that is this removal for all the kernel versions, or there might still be some maintained versions using this driver (e.g., from a quick check it seems the longterm 5.15.91 still has the driver)? If so, do you think this is a valid issue that we need to fix in some prior affected longterm kernel versions? Thank you for your help and comments!

Best,
Hang

Quoted 46 lines · Hang Zhang

__adpt_reset() invokes adpt_hba_reset(), which can free "pHba" on error paths and return an negative error code in those situations. The problem is that "pHba" is from the global pointer "cmd->device->host->hostdata[0]", regardless of the possible free of "pHba", that original global pointer is never nullified and thus may become a dangling pointer and be dereferenced later, potentially causing a use-after-free.

Fix the issue by nullifying "cmd->device->host->hostdata[0]" if adpt_hba_reset() returns a negative error code to __adpt_reset(), which indicates the free of "pHba". Also add a NULL check before any dereference of "pHba", or the aliased global pointer. Note that the similar NULL check already exists at other places, like in adpt_queue_lck().

---
 drivers/scsi/dpt_i2o.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/scsi/dpt_i2o.c +4 -0
index 2e9155ba7408..9827517a1898 100644
@@ -753,6 +753,9 @@ static int __adpt_reset(struct scsi_cmnd* cmd)
  char name[32];
 
  pHba = (adpt_hba*)cmd->device->host->hostdata[0];
+ if (!pHba) {
+ return FAILED;
+ }
  strncpy(name, pHba->name, sizeof(name));
  printk(KERN_WARNING"%s: Hba Reset: scsi id %d: tid: %d\n", name, cmd->device->channel, pHba->channel[cmd->device->channel].tid);
  rcode =  adpt_hba_reset(pHba);
@@ -760,6 +763,7 @@ static int __adpt_reset(struct scsi_cmnd* cmd)
  printk(KERN_WARNING"%s: HBA reset complete\n", name);
  return SUCCESS;
  } else {
+ cmd->device->host->hostdata[0] = NULL;
  printk(KERN_WARNING"%s: HBA reset failed (%x)\n", name, rcode);
  return FAILED;
  }

A message that is not in the archive (sent off-list or not yet received): [email protected]

author
#5 lore raw
Quoted 20 lines · Tomas Henzl

On 2/3/23 21:05, Hang Zhang wrote:

On Fri, Feb 3, 2023 at 11:07 AM Tomas Henzl <[email protected]> wrote:

Hi Hang,

with b04e75a4a8a8 scsi: dpt_i2o: Remove obsolete driver was the driver removed from kernel.

Regards,
Tomas

Hi Tomas, thank you very much for pointing this out! I'm wondering that is this removal for all the kernel versions, or there might still be some maintained versions using this driver (e.g., from a quick check it seems the longterm 5.15.91 still has the driver)? If so, do you think this is a valid issue that we need to fix in some prior affected longterm kernel versions? Thank you for your help and comments!

Quoted 7 lines

It is now removed from the mainline, it can stay elsewhere
for an uncertain time.
If there is a mechanism how to fix patches only in those longterm kernels I don't know. Ask in related mailing lists.
I think that when the driver is now eol and this issue hasn't been noticed before, it's likely not worth the effort to fix it now.

Hi Tomas, thank you for the reply. We will try to gather more information about this. Thanks!

Best,
Hang

Quoted 51 lines

Best,
Hang

On 12/27/22 01:10, Hang Zhang wrote:

__adpt_reset() invokes adpt_hba_reset(), which can free "pHba" on error paths and return an negative error code in those situations. The problem is that "pHba" is from the global pointer "cmd->device->host->hostdata[0]", regardless of the possible free of "pHba", that original global pointer is never nullified and thus may become a dangling pointer and be dereferenced later, potentially causing a use-after-free.

Fix the issue by nullifying "cmd->device->host->hostdata[0]" if adpt_hba_reset() returns a negative error code to __adpt_reset(), which indicates the free of "pHba". Also add a NULL check before any dereference of "pHba", or the aliased global pointer. Note that the similar NULL check already exists at other places, like in adpt_queue_lck().

---
 drivers/scsi/dpt_i2o.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/scsi/dpt_i2o.c +4 -0
index 2e9155ba7408..9827517a1898 100644
@@ -753,6 +753,9 @@ static int __adpt_reset(struct scsi_cmnd* cmd)
  char name[32];
 
  pHba = (adpt_hba*)cmd->device->host->hostdata[0];
+ if (!pHba) {
+ return FAILED;
+ }
  strncpy(name, pHba->name, sizeof(name));
  printk(KERN_WARNING"%s: Hba Reset: scsi id %d: tid: %d\n", name, cmd->device->channel, pHba->channel[cmd->device->channel].tid);
  rcode =  adpt_hba_reset(pHba);
@@ -760,6 +763,7 @@ static int __adpt_reset(struct scsi_cmnd* cmd)
  printk(KERN_WARNING"%s: HBA reset complete\n", name);
  return SUCCESS;
  } else {
+ cmd->device->host->hostdata[0] = NULL;
  printk(KERN_WARNING"%s: HBA reset failed (%x)\n", name, rcode);
  return FAILED;
  }